What Makes a Language Look Like Itself?

Simple statistics can reveal the visual fingerprints of different languages by identifying unique character sequences. A likelihood ratio is used to measure how much more likely a character pattern is to appear in one language versus all others, with additive smoothing applied to handle rare patterns. The analysis was conducted on the top 5,000 words from 20 European languages, extracting all substrings from one to five characters long. The results highlight that unique characters (like Icelandic's 'ð') and common combinations (like Dutch's 'ijk') serve as strong identifiers for their respective languages.